JP7509450B2 - Ranging system and method for reducing transition effects in multi-range material measurements - Patents.com - Google Patents

Description

CROSS-REFERENCE TO RELATED APPLICATIONS
This application is a continuation of U.S. Provisional Patent Application No. 63/016,747, filed April 28, 2020, to Fortney, entitled "ADVANCED ANALOG-TO-DIGITAL CONVERSION SYSTEMS AND METHODS," and U.S. Provisional Patent Application No. 63/034,052, filed June 3, 2020, to Fortney, entitled "ADVANCED DIGITAL-TO-ANALOG SIGNAL GENERATION SYSTEMS AND METHODS," each of which is incorporated herein by reference in its entirety. No. 63/057,745, filed July 28, 2020, to Fortney, entitled "SYNCHRONOUS SOURCE MEASURE SYSTEMS AND METHODS."

This disclosure relates to measurement systems and methods. More specifically, it relates to avoiding glitches or errors caused by changes in ranging of electronics used in the measurements. More generally, it relates to electronics, analytical instrumentation, software, and infrastructure for signal delivery and signal measurement. This disclosure also relates to systems for measuring signals for materials and device characterization and other applications under difficult experimental conditions that can cause high levels of noise and interference.

Materials and device property measurements (e.g., electronic transport properties such as hole, mobility, and carrier concentrations) often require continuous measurements over decades or orders of magnitude of change in the property. Capturing this requires switching from one set of analysis electronics to another, with the different electronics configured for different ranges (e.g., decades or orders of magnitude) in the property being measured. This switching causes glitches and/or gaps in the measurement signal. This also disrupts the data collection process in other ways, such as by causing transients that can corrupt the measurement.

Analog-to-digital converters (ADCs) play a key role in the electronics that amplify, filter, sample, and digitize the measurement signals in these measurement systems. Therefore, the ADC signal processing must be carefully configured for the operating conditions, including the range of the property being measured. However, carefully configuring the ADC system for one range is likely to make it inappropriate for others. This can lead to errors, especially when the property varies across the range. Selective amplification can address these errors. However, the amplifier introduces its own errors. These errors come from amplifier noise, offsets, gain errors, and phase mismatch. In addition, carefully configuring the gain over several ranges requires flexibility that most amplification systems lack. Small signals require large gain to increase resolution and noise performance. As the signal becomes large over the course of the measurement, that same large gain can saturate the ADC. This can cause distortion and signal loss.

To allow for greater flexibility in configuring gain, amplifier stages can be switched on and off, or in and out of the signal chain. At any given time, the amplifiers switched on are the ones configured for the current signal range. When the signal enters another range, the system switches to a different amplifier chain configured for the new range. However, glitching and discontinuities in the measurement often appear during the transition.

FIG 1 illustrates its effect on a conventionally ranged measurement. Specifically, FIG 1 illustrates conventional ranging data 104 across a discontinuity D that is introduced when the measurement signal increases from a lower range r1 to a higher range r2 through a transition _tTR . FIG 2 illustrates an example of a conventional ranging setup 120 that may cause the discontinuity D illustrated in FIG 1. The conventional ranging setup 120 includes two gain chains A and B. Gain chain A is dedicated to and configured for the lower range r1 (FIG 1). Specifically, the gains of amplifier G _A and ADC A are both configured for the lower range r1. Gain chain B is dedicated to and configured for the higher range r2 (FIG 2). This means that the gains of amplifier G _B and ADC B are configured for r2.

When the measurement signal is small (i.e., in the lower range r1), the channel selection component 122 of the ranging setup 120 selects gain chain A. As the measurement signal increases toward the higher range r2 and transitions between the ranges at _tTR , the channel selection component 122 engages the electronics gain chain B. In this way, the channel selection component 122 attempts to ensure that the measurement system has gain configured across two different ranges. However, as shown diagrammatically in FIG. 1, the transition _tTR may introduce a discontinuity D in the measurement data. This is because switching between gain chains A and B may introduce transient signals, noise, or glitches resulting from the "warming up" or start-up of the equipment dedicated to measuring across the transitioned-to range.

Discontinuity D results in two types of ranging errors. These errors occur when two ranges (e.g., r1 and r2) have different configured amplifier profiles (A and B, respectively). In the first type of error, the amplifier profile mismatch causes unwanted amplitude discontinuities or swings (ΔV) in the measured output voltage. In the second type, temporal data discontinuities, data flow may be interrupted during the transition between ranges. In FIG. 1, this is manifested in a data gap in the period from _tTR to _tB . Temporal data discontinuities occur when changing ranges involve new electronics being "warmed up" or engaged, specifically, the amplifier associated with profile B. Collecting data is inaccurate or impossible until these transients dissipate. Transients from cooling down or shutting down the amplifier associated with amplifier profile A may also cause delays or glitches in the measurement system.

It is difficult or impossible to configure a setup such as 120 to eliminate discontinuity D. The configuration is limited by the simplicity and lack of variability of the components (G _A , ADC A, G _B , and ADC B). Thus, there is a critical need for new and improved solutions to provide a robust, high quality low noise source or measurement signal even as the measurement signal varies over decades or orders of magnitude. There is a critical need for flexible solutions to provide smoother transitions between ranges that reduce or eliminate discontinuities such as those shown in FIG. 1.

Aspects of the present disclosure include a measurement system including a gain chain configured to amplify an analog input signal, a range selector configured to select a gain between the analog input signal and multiple analog-to-digital converter (ADC) outputs from multiple ADCs, where each ADC output has a path, and the gain of each output path may be configured from multiple gain stages in the gain chain, and a mixer configured to combine the multiple ADC outputs into a single mixed output.

The plurality of ADCs may comprise a first ADC and a second ADC. The step of combining the plurality of ADC outputs may be performed according to mixed output=αE _first +(1−α)E _second , where E _first may be the output of the first ADC, E _second may be the output of the second ADC, and α may be a mixing parameter that ranges from one to zero. The system may comprise two or more ADCs. A first portion of a gain chain may be connected to a first one of the plurality of ADCs, and a second portion of the gain chain may be connected to a second one of the plurality of ADCs. A range selector may select a gain for a first one of the plurality of ADCs from the first portion of the gain chain, and may select a gain for a second one of the plurality of ADCs from the second portion of the gain chain. Each of the gain stages in the gain chain may be connected to each of the plurality of ADCs via one or more switch banks. The range selector may select a first portion of the shared gain stage for a first one of the plurality of ADCs and a second portion of the shared gain stage for a second one of the plurality of ADCs by setting switches in one or more switch banks. The range selector may comprise first and second multiplexers. The first multiplexer may select the first portion of the shared gain stage. The second multiplexer may select the second portion of the shared gain stage.

Selecting a first portion of the shared gain stage may include configuring a gain for a first one of the plurality of ADCs, and selecting a second portion of the shared gain stage may include configuring a gain for a second one of the plurality of ADCs. Configuring the gains for the first and second ones of the plurality of ADCs may include configuring the gains according to at least one range of the input signal. The mixer may be configured to select an output from the first ADC as a single mixed output when the input signal may be within a first range. The mixer may be configured to select an output from the second ADC as a single mixed output when the input signal may be within a second range. The mixer may be configured to select a mixture of outputs from the first and second ADCs as a single mixed output when the input signal may be between the first and second ranges.

The system may maintain the second ADC online during a first transition period when the input signal may be within a first range. The system may maintain the first ADC online during a second period when the input signal may be within a second range. The range selector may be configured to configure a gain for at least one of the first ADC and the second ADC based on an expected range of the input signal. During a hysteresis period, the system may maintain the first ADC offline. The system may maintain the second ADC online. The system may maintain the gain of the second ADC constant. The hysteresis period may be between the first transition period and the second transition period.

The multiple ADC output paths may comprise two ADC output paths that may be independently configured into high-range and low-range paths. The low-range path may have a first gain for converting the analog input signal. The high-range path may have a second gain for converting the analog input signal. The second gain may be lower than the first gain. The paths may comprise a mixing device configured to combine the lower-range output with the higher-range output. The system may comprise a device configured to vary the amount of gain combined from the low-range and high-range paths. The high-range path may be connected to a first gain chain, and the low-range path may be connected to a second gain chain. The system may comprise a selector for selecting a gain stage of the first gain chain for the first gain and a gain stage of the second gain chain for the second gain. The first and second gains may each comprise a gain stage in a gain chain common to the low-range and high-range paths. The gains of each output path may be substantially identical. The mixer may average the output from each path to reduce noise in a single output.

Aspects of the present disclosure may further include amplifying an analog input signal using a gain chain, selecting a gain between the analog input signal and multiple analog-to-digital converter (ADC) outputs from multiple ADCs, where each ADC output has a path and the gain of each output path may be configured with a gain stage in the gain chain, and combining the multiple ADC outputs into a single mixed output.

A first portion of the gain chain may be connected to a first of the plurality of ADCs, and a second portion of the gain chain may be connected to a second of the plurality of ADCs. Each of the gain stages in the gain chain may be connected to each of the plurality of ADCs via one or more switch banks. The method may further include independently configuring the two ADC output paths into high-range and low-range paths. The method may include applying a first gain from the low-range path to convert the analog input signal. The method may include applying a second gain from the high-range path to convert the analog input signal, the second gain being lower than the first gain. The method may include combining a lower-range output with a higher-range output. The method may include varying an amount of combined gain from the high-range path and the low-range path.
The present invention provides, for example, the following:
(Item 1)
1. A measurement system comprising:
a gain chain configured to amplify the analog input signal;
a range selector configured to select a gain between the analog input signal and a plurality of analog-to-digital converter (ADC) outputs from a plurality of ADCs, each ADC output having a path, the gain of each output path being composed of a plurality of gain stages in the gain chain;
a mixer configured to combine the multiple ADC outputs into a single mixed output;
A measurement system comprising:
(Item 2)
the plurality of ADCs includes a first ADC and a second ADC;
Combining the plurality of ADC outputs comprises:
Mixed output = αE _first + (1-α)E _second
Implemented in accordance with
In the formula,
E _first is the output of the first ADC;
E _second is the output of the second ADC;
α is a mixing parameter that ranges from 1 to zero.
2. The system according to item 1.
(Item 3)
3. The system of any one of claims 1 and 2, comprising two or more ADCs.
(Item 4)
a first portion of the gain chain coupled to a first one of the plurality of ADCs and a second portion of the gain chain coupled to a second one of the plurality of ADCs;
the plurality of ADCs comprises at least two types of ADCs;
The system according to any one of items 1 to 3, wherein the system comprises at least one of the following:
(Item 5)
5. The system of claim 4, wherein the range selector selects a gain for a first one of the plurality of ADCs from a first portion of the gain chain and selects a gain for a second one of the plurality of ADCs from a second portion of the gain chain.
(Item 6)
6. The system of any one of claims 1-5, wherein each of the gain stages in the gain chain is connected to each of the plurality of ADCs via one or more switch banks.
(Item 7)
7. The system of claim 6, wherein the range selector selects a first portion of the shared gain stage for a first one of the plurality of ADCs and a second portion of the shared gain stage for a second one of the plurality of ADCs by setting switches in the one or more switch banks.
(Item 8)
the range selector comprises first and second multiplexers;
the first multiplexer selects a first portion of the shared gain stage and the second multiplexer selects a second portion of the shared gain stage.
7. The system according to item 6.
(Item 9)
9. The system of claim 8, wherein selecting a first portion of the shared gain stage includes configuring a gain for a first one of the plurality of ADCs and selecting a second portion of the shared gain stage includes configuring a gain for a second one of the plurality of ADCs.
(Item 10)
10. The system of claim 9, wherein configuring a gain for a first and second one of the plurality of ADCs comprises configuring the gain according to at least one range of the input signal.
(Item 11)
The mixer includes:
selecting an output from a first ADC as the single mixed output when the input signal is within a first range;
selecting an output from a second ADC as the single mixed output when the input signal is within a second range;
selecting a mixture of the outputs from the first and second ADCs as the single mixed output when the input signal is between the first and second ranges;
The system according to any one of items 1 to 10, configured to perform the following:
(Item 12)
The system comprises:
maintaining the second ADC online during a first transition period when the input signal is within the first range;
maintaining the first ADC online during a second period when the input signal is within the second range;
Item 12. The system according to item 11.
(Item 13)
13. The system of claim 12, wherein the range selector is configured to configure a gain for at least one of the first and second ADCs based on an expected range of the input signal.
(Item 14)
During the hysteresis period, the system:
maintaining the first ADC offline;
maintaining said second ADC online;
maintaining the gain of the second ADC constant;
Item 13. The system according to item 12.
(Item 15)
15. The system of claim 14, wherein the hysteresis period is between the first transition period and the second transition period.
(Item 16)
The plurality of ADC output paths include:
Two ADC output paths that can be independently configured into a high range path and a low range path,
the low range path having a first gain for converting the analog input signal;
the high range path has a second gain for converting the analog input signal, the second gain being lower than the first gain;
Two ADC output paths;
a mixing device configured to combine the lower range output with the higher range output;
a device configured to vary the amount of combined gain from the low range path and the high range path;
The system according to any one of items 1 to 15, comprising:
(Item 17)
17. The system of claim 16, wherein the high range path is connected to a first gain chain and the low range path is connected to a second gain chain.
(Item 18)
17. The system of claim 16, further comprising a selector for selecting a gain stage of the first gain chain for the first gain and for selecting a gain stage of the second gain chain for the second gain.
(Item 19)
17. The system of claim 16, wherein the first and second gains each comprise a gain stage in a gain chain common to the low range path and the high range path.
(Item 20)
the gain of each output path being substantially the same;
the mixer averages the output from each path to reduce noise at the single output;
20. The system of any one of items 1-19.
(Item 21)
1. A method comprising:
amplifying an analog input signal using a gain chain;
selecting a gain between the analog input signal and a plurality of analog-to-digital converter (ADC) outputs from a plurality of ADCs, each ADC output having a path, the gain of each output path being configured from a gain stage in the gain chain;
combining said multiple ADC outputs into a single mixed output;
A method comprising:
(Item 22)
22. The method of claim 21, wherein a first portion of the gain chain is connected to a first one of the plurality of ADCs and a second portion of the gain chain is connected to a second one of the plurality of ADCs.
(Item 23)
23. The method of any one of claims 21 and 22, wherein each of the gain stages in the gain chain is connected to each of the plurality of ADCs via one or more switch banks.
(Item 24)
Configuring two ADC output paths independently into a high range path and a low range path;
applying a first gain from the low range path to convert the analog input signal;
applying a second gain from the high range path to convert the analog input signal, the second gain being lower than the first gain; and
combining said lower range output with said higher range output;
Varying the amount of combined gain from the high range path and the low range path;
24. The method according to any one of items 21 to 23, further comprising:

FIG. 1 shows the effect of range transitions on data collected by a typical measurement system without seamless ranging capabilities.

FIG. 2 shows an example of a conventional ranging configuration 120 that may cause the discontinuity D shown in FIG.

FIG. 3 compares a voltage measurement with seamless ranging 302 with the same measurement made with a conventional setup 104 (from FIG. 1) that lacks seamless ranging capabilities.

FIG. 4 is one variation 400 that implements seamless ranging via dual amplification chains.

FIG. 5 illustrates another exemplary amplification chain 500 according to an aspect of the present disclosure.

FIG. 6A provides a schematic example of an automatic ranging algorithm 600 in accordance with an aspect of the present disclosure.

FIG. 6B shows measurement data corresponding to the automatic ranging algorithm 600.

FIG. 6C illustrates the algorithm 600 in flow chart form.

FIG. 6D illustrates another auto ranging algorithm 620 in flow chart form.

FIG. 7A illustrates another variation 700 for sharing gain stages while using multiple ADCs (ie, ADC A 708a and ADC B 708b) in accordance with an aspect of the present disclosure.

FIG. 7B shows that the measurement signal 750a of the variation 700 exhibits a discontinuity 752 in magnitude at the r1/r2 transition in _tTR .

FIG. 8 illustrates another variation 800 in accordance with aspects of the present disclosure that places a preamplifier (preamp) 804a prior to the gain chain 700c and/or a preamplifier 804b in one of the two paths.

FIG. 9 illustrates an interpolation algorithm 910 directed to eliminating or reducing discontinuities 752 in accordance with an aspect of this disclosure.

FIG. 10A illustrates a generalized variation 1000 of gain selection that may be used in accordance with this disclosure.

FIG. 10B shows an example gain path (Gain Path A) that may be created using variation 1000.

FIG. 10C illustrates another gain path (gain path B) of variation 1000 that includes two variations, a high range and a low range variation.

FIG. 11 illustrates another variation 1100 including variable gain selection by gain stage selectors 1116a-1116n in accordance with an aspect of the disclosure.

FIG. 12A is a schematic diagram of an exemplary mixing and auto-ranging algorithm 1200 that may be implemented by range mixers 410 , 510 , 710 , 1010 , and 1110 .

FIG. 12B illustrates an asymmetric auto-ranging algorithm 1250 that may be implemented by range mixers 410 , 510 , 710 , 1010 , and 1110 .

FIG. 13A shows a flowchart 1300 illustrating a range change prediction algorithm 1300 that may be performed by range mixers 410, 510, 710, 1010, and 1110 when implementing the algorithms disclosed herein (eg, 600, 620, 910, 1200, and 1250).

FIG. 13B shows another portion of the flowchart 1300.

FIG. 13C shows another portion of the flowchart 1300.

FIG. 14 shows an example implementation of the range blending algorithm 1400 . FIG. 14 shows an example implementation of the range blending algorithm 1400 .

FIG. 15 illustrates a measurement signal chain 1500 between an example head unit 1550 and an example measurement pod 1560 that may use variations 400, 500, 700, 800, 1000, 1100 and algorithms 600, 620, 910, 1200, 1250, 1300, and 1400.

FIG. 16 illustrates another example variation 1600 in which a head unit 1550 may have six channels that may support three measurement type pods 1560a and three source type pods 1560b in accordance with an aspect of the disclosure.

DETAILED DESCRIPTION The present disclosure introduces systems and methods that can accommodate measurements over a wide dynamic range with relatively little error, noise, or glitching. Here, "glitching" refers to unintended irregularities or inconsistencies that can adversely affect measurements or operations. The variations disclosed herein accomplish this in several different ways. One way is to separately and dynamically configure gain chains for the separate ranges. Another is to string the separate ranges together by blending the gain profiles for the ranges. Yet another is to introduce a shared gain stage that can be dynamically assigned to the separate ranges. These and further methods are generally referred to herein as "seamless ranging." They are discussed in more detail below.

Figure 3 compares a voltage measurement with seamless ranging 302 according to the present disclosure to the same measurement made by a conventional setup 104 lacking seamless ranging capabilities. Figure 3 shows a discontinuity D in the measurement data 104 across the range transition Δt because a different set of devices with different measurement profiles (e.g., accuracy, gain, etc.) are used to measure data in ranges r1 and r2. As discussed, switching between ranges r1 and r2 in system 120 may involve transient signals, noise, or glitches resulting from the "warm-up" or start-up of equipment dedicated to measuring across the transitioned range (r2).

3 also illustrates how the transition Δt can be smoothed by the seamless ranging capabilities described herein (continuous ranging measurement data 302). This smoothing effect is represented in FIG. 3 as the avoidance of discontinuities D by the continuous ranging data 302. Although only two example ranges r1 and r2 are discussed in the context of FIG. 3, it should be understood that the continuous ranging technique can be applied to any suitable number of ranges associated with a particular measurement. For example, the number of ranges may be three, four, or more in some cases. In each of these cases, the continuous ranging can be configured to ensure a smooth transition between each range change, regardless of the direction of the range change (i.e., whether the range change involves an increase as shown in FIG. 3 or a decrease in the measurement (not shown)).

Continuous ranging addresses the two ranges r1 and r2 using separate signal amplification/gain chains that can be applied independently and/or in parallel. As an example, a specific implementation will be discussed below in the context of Figs. 4 and 5. Addressing each range r1 and r2 separately and/or in parallel allows the configuration of amplification chains for the inactive or "cold" ranges (i.e., ranges not currently employed in the measurement, e.g., range r2 when t< _tTR or range r1 when t> _tTR ) on the basis that data is collected by active amplification changes. Keeping the amplification chains for the inactive or cold ranges online in parallel with the active range measurements can avoid start-up transients when the inactive ranges are eventually engaged. This also allows "range mixing," where gain chains per range are applied in combination to promote a smooth change in data over the transition Δt from range r1 to r2 (or vice versa). That is, amplification chains from both ranges can be applied simultaneously to smooth the data over the range transition Δt. This can be done, for example, via a software mixer and/or can then smoothly transition from r1 to r2 and vice versa.

4 is one variation 400 that implements seamless ranging via dual amplification chains. As shown in FIG. 4, the lower gain chain 402 (i.e., the gain chain with lower amplification) and the higher gain chain 404 (i.e., the gain chain with higher amplification) are the same except for 1) different ADCs (408a and 408b, respectively) and 2) an additional amplifier 406 in the higher gain chain 404, which gives it a higher gain than the lower gain chain 402. The outputs from the ADCs 408a and 408b are combined by a mixer 410 and used in the acquisition routine of the measurement pod 104 for ranging measurements. In the chain 400, the combination can be weighted by a factor α. The factor α can be dynamically chosen to ensure a smooth transition over the ranging transition Δt (e.g., using range mixing to avoid discontinuity D of FIG. 3). The coefficient α can be set by a user, but it is often set by a ranging algorithm (e.g., algorithms 600, 650, 910, 1200, 1250, 1300, and 1400, discussed in more detail below).

Figure 5 shows another example amplification chain 500 used in seamless ranging. Chain 500 includes a lower gain section 502 and a higher gain section 504 that are identical except for 1) different ADCs (508a and 508b, respectively), 2) an additional amplifier 506 in the higher gain section 504, which gives it a higher gain than the lower gain section 502, and 3) the lower gain section 502 and the higher gain section 504 are connected to a gain stage 512 via muxes 514a and 514b, respectively.

As shown in FIG. 5, the amplification provided from gain stages 512a and 512b to the lower and higher gain sections 502 and 504 can be selected via muxes 514a and 514b, respectively. In this way, chain 500 may use fewer dedicated amplifiers than chain 400 and provide the combination to mixer 510. Using identical gain stages 512a and 512b (and amplifiers) for the lower and higher gain sections 502 and 504 is not only more efficient. It also introduces less noise in the system that may result from glitches or mismatches between different amplifiers. For example, each amplifier may have transients that are avoided when they are used all the time, regardless of the range to which they are applied. Any idiosyncrasies, ranging problems, or errors caused by amplifiers 512a and 512b will be present in all ranges to which they are applied. As discussed in more detail below, this can ensure consistency and smoothness of the overall trend and behavior of the measurement data, which is often a more important aspect in materials measurement than precisely measured amplitude.

As in chain 400, the combination in chain 500, blend 510, can be weighted by a factor α. The factor α can be dynamically chosen to ensure a smooth transition across the ranging transition Δt (e.g., using range blending to avoid discontinuity D of FIG. 3). The factor α can be set by the user, but is often set by a ranging algorithm (e.g., algorithm 600 shown in FIG. 6A). α can be set by any of the methods described herein for gain, chain, or signal blending.

In variations including chains 400 and 500, and others, seamless ranging may include automatic ranging. FIG. 6A provides an illustration of an automatic ranging algorithm 600 that may be used, for example, in conjunction with seamless ranging in variations 400 and 500. FIG. 6C illustrates algorithm 600 in flow chart form.

Algorithm 600 varies the ranges as a measurement signal 650, shown in Figure 6B, varies incrementally from ranges r1, r2, and r3. Signal 650 transitions from range r1 to r2 at t = _tTR(1-2) and from range r2 to r3 at tTR _(2-3) . Figure 6A shows the response of algorithm 600 in terms of applying dedicated gain chains in ranges r1, r2, and r3 across those transitions.

As shown in Figures 6A and 6C, algorithm 600 provides the configured gain for 100% r1 (e.g., draws from the higher gain portion 404 in chain 400 of Figure 4 and provides a higher gain to the lower of the two ranges) for a period 602 prior to the transition from r1 to r2 (tTR(1-2 ₎ ). Figures 6A and 6C also show that algorithm 600 blends the gain profiles for r1 and r2 (e.g., draws from the higher gain portion 404 and the lower gain portion 404) as the measurement signal approaches the transition _tTR(1-2) . The r1/r2 blending period prior to this transition is labeled 604. As discussed above, the blending avoids glitches and/or gaps in the data during the r1/r2 range transition. After the r1/r2 transition at t _TR(1-2) , the algorithm 600 applies the r2 gain without blending (e.g., drawing from the lower gain portion 402 in the chain 400 of FIG. 4). Figures 6A, 6B, and 6C show that the algorithm similarly changes from r2 to r3 at t _TR(2-3) , i.e., first blending the gain profiles for r2 and r3 during period 608, and then providing only the r3 configuration gain during period 610.

6A also shows a region of hysteresis 612 during period 606 (r2 only). During hysteresis 612, there is no anticipated ranging (i.e., only one gain portion of the gain chain is active, in this case the gain chain for r2). The applied gain during hysteresis may also be constant. This avoids alternating switching between ranges due to noise or signal fluctuations. Once the measurement signal 650 moves closer to r3, the hysteresis period 612 ends. Period 614 represents a period in which the range change from r2 to r3 is anticipated by engaging the gain chain for r3 (not shown). The gain chain corresponding to r3 is engaged during 614 for both calibration and transient avoidance purposes as discussed above. It should be understood that no hysteresis or anticipation of the upper portion of the range is shown for the r1/r2 transition, but they may be applied to that transition as well.

6A illustrates the operation of algorithm 600 as the measurement signal increases, it should be understood that the algorithm equally applies as the measurement signal decreases (e.g., from a higher range r3 to a lower range r2 to the lowest range r1). This is illustrated via flow chart 620 in FIG. 6D. In this case, rather than predicting a period above the range, algorithm 600 would predict a period below the range (e.g., transitioning downward from r3 to r2 at _tTR(3-2) (step 624 of chart 620), etc.).

While Figures 6A, 6B, and 6C show algorithms 600 and 620 that handle range changes between three example ranges r1, r2, and r3, it should be understood that this may handle range changes between any number of ranges suitable for experimentation in the same manner. Other variations of algorithms 600 and 620 may include many other algorithms and/or range/parameter settings and any suitable number of range transitions.

The mixers 410 and 510 in the chains 400 and 500, respectively, can operate according to any suitable mixing algorithm to achieve the smoothing effect (302) shown in Figure 3. The mixers 401 and 501 may be digital. They may not have any need for independent calibration. In one variant, the mixed output of 410 and 510 may be controlled by an algorithm similar to the following:
Output signal V (for mixer 402 or 502)=αE _A +(1−α)E _B (1)
In the formula,
E _A is the output of the first ADC (ADC A 408a or ADC A 508a);
_EB is the output of the second ADC (ADC A 408b or ADC A 508b);
α is a blending parameter that can range, for example, from one to zero.

It should be understood that Equation 1 is not the only mixing algorithm that may be applied by mixers 410 and 510. For example, the mixer may simply average the output of each path to reduce noise. Equation 1 applies linear weighting (α) to the contributions of E _A and E _B. However, non-linear weighting is also contemplated and should be considered within the scope of this disclosure. In practice, the weighting may include any suitable mathematical form. Examples include, but are not limited to, quadratic polynomials, cubic polynomials, and any suitable polynomial. Exponential and logarithmic functions, and differential equations are all contemplated within the scope of this disclosure.

The exact form of the weighting or mixing function should depend on factors such as the gains of the various amplifiers in the system (e.g., amplifiers in 402, 404, and 512 and amplifier 506) and other components such as ADCs (e.g., ADCs 408a, 408b, 508a, and 508b). This may also depend on the details of the mixers 401 and 510 used in the circuit. This may depend on the following example characteristics of these components: frequency response, gain value, nonlinearity, sensitivity to variations in the input. In addition, the parameter α need not vary from one to zero, as in the above example. The parameter α and any other values adopted by the mixers 410 and 510 may depend on the details of the gain stages and gains 506 in the chains 402, 404, and 512. This may include any suitable value to balance the gains and eliminate or reduce the discontinuity D (FIG. 3).

Figure 7A shows another variation 700 that shares gain stages while using multiple ADCs (i.e., ADC A 708a and ADC B 708b). Figure 7A shows the architecture of the chain 700 itself. Figure 7B is a plot 750 comparing the response 750a of the chain 700 to a conventional ranging system of the prior art, such as 120.

Chain 700 includes two signal paths 700a and 700b, each including an ADC (ADC A 708a and ADC B 708b, respectively) and a multiplexer (mux 706a and 706b, respectively). Multiplexers 706a and 706b select from gain stages 704a-704c from gain chain 700c. Thus, signal paths 700a and 700b have independently configurable gains based on their selection.

Each independently configurable gain delivered to paths 700a and 700b can be any combination of the outputs of amplifiers 704a, 704b, and 704c with gains A1, A2, and A3, respectively. Gains A1, A2, and A3 can be 1, any suitable positive value greater than 1, and any suitable negative value with an absolute value greater than 1, respectively. The gains can be selected for any reason and based on any criteria, but they are typically selected by muxes 706a and 706b based on the range of the input signal 702 to best accommodate that signal. It should be understood that many different combinations are possible and are within the scope of this disclosure.

For example, input signal 702 may be in a range that is best amplified by the combined gain from gain stages 704a and 704b (i.e., a gain equal to the product of A1 and A2) and ADC A 708a. This range may correspond, for example, to the lower range r1 of FIG. 6A, which requires a relatively high gain. In this case, mux 706a would select input 707a and send that gain to ADC A 708a. After processing in ADC A 708a, the signal is sent to mixer 710. In this case, mixer 710 would select only ADC A 708a input, since ADC A 708a is in the appropriate range and signal (e.g., set α in Equation 1 equal to 1). At the same time, mux 706b may be set so that ADC 708b has a configured gain for the upper range r2. This may be a lower gain for the lower range r1 corresponding to a higher signal amplification. Just as an example, this lower gain may be A1. In this example, ADC B 708b is not used to generate output signal 712 as long as input 702 is within range r1. Typically, paths 700b and their associated unused ranges would be considered cold since they are not actively providing an output to mixer 710. However, even while cold, paths 700b may still be operational to avoid transient events that occur during turn-on or warm-up.

If the input signal 702 is increasing and therefore at risk of saturating ADC A 708a by approaching transition _tTR , then ADC B 708b (the "cold" range) may be engaged. The path associated with ADC B 708b can be set to a higher range (lower gain). For example, ADC B 708b could be fed the output of gain stage A1 (704a), which would result in ADC B 708b being in a higher range (lower gain) than ADC A 708a in path 700a.

While the input signal 702 is at a desired level for ADC A 708a, the mixer 710 is set to receive only the contribution of ADC A 708a at output 712. This corresponds to the lower range r1 in FIG. 3 away from the transition point _tTR . However, as the input signal 702 increases towards _tTR (and the upper range r2 where ADC B 708b is configured), it becomes more advantageous for ADC B 708b to take over. Before the transition _tTR , ADC 708b "warms up" by starting to measure the input signal 702. In this configuration, which corresponds to step 614 in FIGS. 6A and 6C, the mixer 710 is still set to send only the signal from ADC A 708a to output 712. Once the transient in ADC B 708b processing of the input signal 702 has subsided, mixer 710 begins to provide a signal at output 712 that is a combination of the outputs from ADC A 708 and ADC B 708b. This mixed output may be, for example, according to Equation 1. Mixer 710 gradually increases the contribution from ADC B 708b until the system is well within range r2. At that point, which corresponds to step 606 in Figures 6A and 6C, mixer 710 may block or eliminate the contribution from ADC A 708a. This is because ADC B 708b is configured for r2. In the exemplary case, mux 706b is set so that ADC B 708b receives a lower gain (only A1, as opposed to the product of A1 and A2). This corresponds to input 707b. The generation of output 712 by mixing the signals from the two ADC paths 700a and 700b to smooth the transition across D (FIG. 6A) is seamless ranging.

In this scenario, the singularity and/or error associated with gain stage 704a (gain of A1) is common between measurement signal 750a (FIG. 7B) of both ranges r1 and r2 and their range paths 700a and 700b, respectively. Thus, the transition _tTR between the ranges has gain commonality. This reduces the discrepancy between the ranges. The effect on the measurement data is shown diagrammatically in FIG. 7B. Specifically, FIG. 7B shows how the signal 750a measured by 700 is more similar in ranges r1 and r2 (parts A and B, respectively) than the same output measured by the prior art configuration (e.g., 100 shown in FIG. 1). In other words, the measurement signals 750a are more similar when they are measuring the same signal (seamless ranging) compared to a completely different set of gains in each range (prior art). In FIG. 7B, both the prior art system and the seamless ranging system 750a have the same output for part A (range r1).

Once the input signal 702b is passed to ADC B 708b, ADC A 708a is now cold. Even while cold, the gain of ADC A 708a remains configured to anticipate where the signal will go next. ADC A 708a may, for example, stay within the r2 configuration range to anticipate a return to that range. Alternatively, ADC A 708a may change its range by resetting mux 706b for a different gain. ADC A 708a may do this in anticipation of the signal continuing to increase or decrease, depending on the initial conditions of each signal path.

As discussed in the context of FIG. 6D, the transition described above may be performed in reverse for a decreasing input signal 702. In other words, if the signal 702 is decreasing from range r2 to r1, the mixer will initially be set to feed only the contribution from path 700b to the output 712. This is because ADC B 708b is configured for range r2 by setting mux 708a to receive input 707b (lower gain A1). As the input 702 decreases toward _tTR , ADC 708a is turned on to warm up and allow transients to die out. At this stage, the mixer 712 is still set so that the output 712 receives only the 708b contribution. Once the input 702 approaches _tTR , the mixer 710 is set to combine the 700a and 700b contributions and create a seamless transition. As the input 702 decreases beyond _tTR to r1, the mixer 710 is reset so that only the configured path for r1 (i.e., 700a, including ADC A 708a) contributes to the output 712. As discussed in the example above, this gain may be the product of A1 and A2, set by mux 706a.

In a system with many gain stages such as 700, the input signal 702 can be passed alternately between ADCs 708a and 708b as the input signal increases or decreases. Each time, the cold ADC will predict the range required for the changing signal as explained above. During this process, the gain can be changed for the cold ADC while the output is taken from the active ADC. This results in a constant output within the desired range, as shown in FIG. 7B, and reduced discrepancies due to gain variations within each range.

7B shows that the measurement signal 750a exhibits a discontinuity 752 in magnitude at the r1/r2 transition at _tTR . This is merely for illustrative purposes and may not be present in all implementations. The discontinuity 752 results from a situation where the gains applied to paths 700a and 700b, configured for each range r1/r2, are slightly mismatched at the transition _tTR . In many variations, it may be possible to adjust the gains for each path 700a and 700b to eliminate the discontinuity 752. However, it may be more important to configure the gains to best represent their individual ranges. In this case, the discontinuity 752 is a known artifact of the measurement electronics and can be addressed in several ways (e.g., by curve fitting/smoothing, etc.) in post-processing of the measurement data 750a.

FIG. 8 shows another variation 800 that places a preamplifier (preamp) 804a prior to the gain chain 700c and/or a preamplifier 804b in one of the two paths. FIG. 8 shows a preamplifier 804b in the path 700b associated with ADC B 708b. However, it should be understood that a preamplifier 804b could also be placed in a similar position in the path 700a associated with ADC A 708a. Except for the addition of preamplifiers 804a and 804b, variation 800 is the same as variation 700 of FIG. 7A.

Preamplifiers 804a and 804b can provide several benefits to variation 800. For example, preamplifier 804a can buffer input signal 702 from other components in variation 800. This can be advantageous because directly connecting input 702 to multiple buffers or switching elements degrades performance. These elements often impart bias currents and switching capacitance to input 702. Preamplifier 804b can be placed in the path (either 700a or 700b) that is typically associated with the range that requires extra gain. This can be, for example, the lowest range (e.g., range r1 in FIG. 6A). Having an extra gain stage "wired" into one of the paths makes it simpler and easier to apply the appropriate gain to that path.

Figure 9 shows an interpolation algorithm 910 directed to eliminating or reducing the discontinuity 752. The algorithm 910 may be implemented by the mixer 710 (Figures 7 and 8) for both paths 700a and 700b.

In some applications, particularly in materials research, the discontinuity 752 itself may be a bigger problem than other quantitative error sources. This is especially true when the overall characteristics of the measurement signal 750a are most important for describing material properties, not its precisely measured value. In many cases, the measurements may be assessed in relative or normalized terms to emphasize the behavior over precise amplitudes. In these cases, the mixer 710 may interpolate its two inputs from the ADCs 708a and 708b to maintain a smooth transition between the ranges r1 and r2. Such an interpolation may be performed via Equation 1. This may also be performed using another suitable mathematical or signal processing means for interpolating the signals from the ADCs 708a and 708b. As shown in FIG. 9, the interpolation 910 is typically performed only over a period 920 proximate to the transition time _tTR . The period 920 may correspond, for example, to Δt shown in FIG. 3. However, it should be understood that the interpolation 910 need not be limited to any particular period. Because the contribution of each of the signals from the two ADCs 708a and 708b is variable, an interpolation 910 may be performed throughout the measurement.

FIG. 10A shows a variation 1000 that includes additional freedom in terms of gain with selection for each path 1000a and 1000b associated with multiple amplifiers 1004a-1004n in a common gain chain 1000c. Gain stage selection can be performed in variation 1000 by two series of switch banks 1006a and 1006b. Each bank includes a switch, e.g., switch 1014a, that can connect or disconnect a data converter (e.g., ADC) 1008a or 1008b to or from each gain stage in the chain 1000c. This is how each amplifier 1004a-1004n can be connected independently.

It should be understood that the switch banks 1006a and 1006b may be implemented in a number of suitable ways. Solid state switching may be used. Alternatively, mechanical relay switching may be used. Any other suitable switching or connection method may be used. The individual switches (e.g., 1014a) may be present and operated individually. Alternatively, they may be operated as part of an integrated circuit or other integrated device. They may be triggered by any suitable means, including user input, any of the algorithms described herein (e.g., algorithms 600, 620, and 910, etc.). The switch banks 1006a and 1006b may also be dynamically operated such that the gains fed to the switching and data converters 1008a and 1008b may be dynamically changed (e.g., at any time within the ranges r1 and r2 of FIG. 6A).

Figure 10A illustrates how a gain path can be created that uses a common gain chain to amplify an input signal. The points before and after each gain stage 1014a-1014n in the common gain chain 1000c can be selected by a number of ranges. In Figure 10A, switching means 1006a and 1006b and/or a controller can be used to select a point on the common gain chain 1000c to pass the input signal to either the top data converter 1008a or the bottom data converter 1008b.

As shown in FIG. 10A, mixer 1010 selects or mixes the outputs from data converters 1008a and 1008b and feeds them to data output 1012. Mixer 1010 may operate in a similar or identical manner to mixers 410, 510, and 710. For example, mixer 1010 may use Equation 1 to mix 1008a and 1008b outputs. It may do so based on any information used by mixers 410, 510, and 710 (e.g., user input, algorithm 600, etc.).

Although FIG. 10A shows only two converters, it should be understood that variation 1000 (and variations 500, 700, and 800) may be used with any suitable number of data converters. One exemplary configuration is to assign a data converter for each independent range. Thus, for example, if a measurement includes four ranges r1-r4, four independent data converters may be used.

FIG. 10A shows that variation 1000 includes any number (n) of gain stages 1004-1004n in gain chain 1000c. Generally, the more gain stages included in 1000c, the more flexibility there is in allowing data converters 1008a and 1008b to represent a particular range. In some variations, such as variation 1000, n is two or more times greater than the number of data converters 1008m.

Although FIG. 10A shows gain stages 1004a-1004n appearing to be of the same or similar type, this may not necessarily be the case. In variations, it may be advantageous to use different types of gain stages with different gains. The benefit of having a common gain chain 1000c is that fewer portions of the system need to be calibrated. In conventional systems, two completely independent gain paths needed to be calibrated. In the present disclosure, the gain stages can be calibrated independently of range. This can reduce the time it takes to calibrate the entire system.

Typically, most or all of the gain stages in the chain 1000c are active. In some cases, it may be useful to deactivate (e.g., generate active or expected ranges) some gain stages 1004a-1004n while not in use. For example, some types of gain stages 1004a-1004n may not handle saturation well without producing errors. In that case, such gain stages would be advantageously deactivated once a risk of saturation is detected. Doing so may allow for faster transitions (i.e., by only activating a range when that range is capable of amplifying the signal adequately). Unused ranges 1004a-1004n may also be deactivated to reduce power draw, heat generation, etc.

10B shows an example gain path (gain path A) that may be created using variation 1000. To create gain path A, switch 1014c is engaged. This causes gain path A to be amplified by gain stages 1004a and 1004b (without any other gain stages). The amplified signal is then sent to data converter 1008a. The path is then mixed with another path by mixer 1010 and sent to data output 1012. In a variation, mixer 1010 may send only the signal from gain path A to data output 1012. In others, it may mix the paths by any of the means or algorithms disclosed herein (e.g., equations, algorithm 600, etc.).

FIG. 10C shows another gain path (gain path B) that includes two variations, a high range and a low range variation. Both the high and low variations use data converter 1008b instead of converter 1008a. Thus, gain path B can be separately and independently engaged with gain path A. Gain paths A and B can be mixed together by mixer 1010 to form data output 1012.

The higher range path of gain path B contains less gain and may be more appropriate for higher ranges (e.g., r2 in FIG. 6A). It does so by triggering switch 1014f, which causes the path to contain gain from only one stage, namely, 1004a. The lower range path is obtained by triggering switch 1014h while switch 1014f is not triggered. The lower range path contains two extra gain stages, namely, 1004b and 1004c, along with gain stage 1004a. This gives it a much higher gain that may be more appropriate for lower ranges (e.g., r1 in FIG. 6A).

Variation 1000 may switch between any of these gain paths as needed. It may do so, for example, according to any of algorithms 600, 620, and 910. For example, variation 1000 may use gain path A first, since gain path A is the lowest gain. This may simultaneously bring gain path B online, warming it up and removing transients. In this scenario, gain path B would be in its lower range configuration, anticipating that the measurement signal would use it first since it is increasing from the lower range (i.e., the lower range associated with gain path A). As the measurement signal continues to increase, mixer 1010 may mix gain paths A and B, with gain path B in the lower range configuration. As the measurement signal continues to increase, mixer 1010 may send only gain path B to data output 1012. As the signal continues to increase beyond this point, the higher range configuration of gain path B may be triggered by turning off 1014h and turning on 1014f. This will give the input signal 1002 the minimum amount of gain (i.e., only the gain from gain stage 1004a) that corresponds to being in the highest range.

FIG. 11 shows another variation 1100 that includes variable gain selection by another means, namely, gain stage selectors 1116a-1116n. Variation 1100 selects gain from among two stages 1104a and 1104b. However, it should be understood that this is merely exemplary. Any suitable number n of gain stages 1104 may be included in 1100.

In variation 1000, each data converter 1108a-1108n is connected to its own gain stage selector 1116a-1116n. However, other configurations in which data converters 1108 share a gain stage selector 1116 are also possible.

Variation 1100 includes a number n of data converters that can be large. In general, the number n of converters can be chosen so that there is one converter per range. In other situations, it may be advantageous to include either more or fewer converters than ranges. Multiple ranges/gain stages are also useful, for example, in pulsed input signal measurement applications. If an input signal transitions through multiple ranges, it may be useful to measure the pulse across several ranges with different gains. It should be understood that any suitable number of gain stages, greater or less than n, may be used.

It may be useful to have a range that always measures at some point in a common gain chain while other ranges alternately pass the signal between desired gains. One variation could use a low-cost ADC to initialize the input signal at a low gain and use this information to quickly configure the gain in a high-quality ADC. This could be beneficial for inputs that vary between different sources. Input signals with large amplitude spikes could also cause problems with the measurement system. Thus, by having multiple ADCs measuring simultaneously, accurate measurements could be achieved when the input is within its "normal" range but still be able to measure the signal spikes. In another variation, the input could benefit from using different types of ADCs simultaneously to measure the signal. A high-speed ADC along with a high-resolution ADC would allow different types of signals to be measured and converted without sacrificing performance. All of these variations could use predictive algorithms along with other ADCs that measure the input signal for other purposes. Many communication signals exhibit this type of signal characteristic.

As shown in FIG. 11, variation 1100 includes a range mixer 1110. Range mixer 1110 mixes the outputs of data converters 1108a-1108n and provides to data output 1112. Range mixer 1110 can mix the outputs according to any of the methods disclosed herein in the context of other range mixers (e.g., in the context of range mixer 1010). Many different types of mixing algorithms can also be designed to combine different ranges and more accurately measure the varying input signal.

As shown in FIG. 11, each gain stage selector 1116a-1116n can provide any combination of gain stages 1104a and 1104b to data converters 1108a-1108n. The combination can be selected by any of the means of gain selection disclosed herein, including user input, any of the algorithms disclosed herein (e.g., 600, 620, and 910).

Figure 12A is a schematic diagram of an exemplary mixing and auto-ranging algorithm 1200 that may be implemented by range mixers 410, 510, 710, 1010, and 1110. The algorithm 1200 mixes three ranges A1, A2, and A3 as shown in Figure 12A. For purely illustrative purposes, A1>A2>A3. It should be understood that higher gains are typically associated with lower ranges in the measured variable, and vice versa. Thus, the exemplary gain configuration of A1>A2>A3 would most likely correspond to the following measurement range configuration: r1<r2<r3. In this situation, the highest gain A1 would be applied to the lowest range in the measurement data r1, and so on. Figure 12A shows the gain increasing (from top to bottom) from A3 to A1 as the value of the measurement signal decreases, i.e., as the measurement signal decreases within the range from r3 to r1.

When the measurement signal is in the highest range (e.g., range r3 in FIG. 6A), the algorithm 1200 applies the lowest gain A3. As the measurement signal decreases and approaches the next lowest range, i.e., the range where the next higher gain A2 is desired, the mixers (e.g., 410, 510, 710, 1010, and 1110) become active. This occurs in stage 1204. In stage 1204, the mixers combine A3 and A2 to smooth the transition. In 1206, the transition between the A3 and A2 ranges is complete. The mixers apply only A2. In stage 1208, the measurement data is definitely in the A2 range. Now, any switching between stages or mixing by the mixer would be an error. Therefore, the algorithm 1200 applies hysteresis to prevent changes in the expected range. This ensures that there is no false switching of the electronics based on noise or aberrations in the data. In step 1210, the measurement data is further reduced and approaches the lowest range in the measurement (e.g., r1 in FIG. 6A) where the highest gain A1 is most appropriate. Thus, the algorithm 1200 "warms up" the A1 gain profile. The mixer does not actually engage the A1 gain on the measurement signal at this point. Instead, it is switched on to remove any transients that may occur. The mixer begins to actively mix the A2 and A1 ranges in step 1212. This is because the measurement signal is now close enough to the highest gain A1/lowest measurement range r1 to smooth the transition. Finally, in step 1214, the measurement data is now firmly in the A1 range. The mixer provides only the A1 gain.

Although FIG. 12A has been described in terms of an increase in gain from the lowest A3 gain to the highest A1 (a decrease in range from the highest measurement range r3 to the lowest measurement range r1), it should be understood that FIG. 12A is bidirectional. That is, algorithm 1200 can also proceed where gain decreases from A1 to A3, which corresponds to an increase in the range of the measurement data from r1 to r3. In that case, algorithm 1200 would follow steps in the reverse order, i.e., 1214-1202.

The auto ranging algorithm may be different for any given application and need not be symmetric or linear as shown in FIG. 12A. An asymmetric variation 1250 is shown in FIG. 12B. In FIG. 12B, there are three ranges defined by the orders of magnitude of the measurement data: 10, 1, and 0.1. Note that the numbers 10, 1, and 0.1 refer to the orders of magnitude of the range in the variable being measured (e.g., voltage). This differs from FIG. 12A, where the ranges are referenced by their gains A1, A2, and A3. Since the gain is inversely proportional to the variable being measured, the lowest measurement range 0.1 corresponds to the highest gain ( _A0.1 ). The highest measurement range 10 corresponds to the lowest gain ( _A10 ). Since a change between the 10 and 0.1 ranges represents a two order of magnitude change in the measurement data, great care must be taken in range blending. Since the measurement signal is very small, especially within the 0.1 range, this can easily be overwhelmed by range blending. Therefore, algorithm 1250 applies range blending judiciously.

When the measurement signal is within the highest measurement data range 10, the algorithm 1250 applies the lowest gain appropriate for that range (A ₁₀ ). This is stage 1252 in FIG. 12. As the measurement signal decreases and approaches the next highest measurement data range, 1, the mixer becomes active. This occurs in stage 1254. In stage 1254, the mixer combines the gains for the 10 (A ₁₀ ) and 1 (A ₁ ) ranges to smooth the transition. In 1256, the transition between the 10 and 1 ranges is complete. The mixer applies only the gain for 1 (A ₁ ). However, the difference in range between the 1 and 10 ranges is so large that the electronics for the 10 (A ₁₀ ) measurement data range remain warmed up. There is no mixing, but the mixer is ready to switch ranges as needed to prevent saturation. At stage 1258, the measurement data is firmly within the 1 range, so switching to range 10 is not possible. Now, any switching between stages or mixing by the mixer would be incorrect. Therefore, the algorithm 1250 applies hysteresis to prevent expected range changes. This ensures that there is no incorrect switching of the electronics based on noise or aberrations in the data. At step 1260, the measurement data decreases enough to approach the lowest 0.1 measurement data range. At this step, the mixer anticipates the downward range change by engaging the electronics for the 0.1 range but keeping them offline (i.e., not mixing ranges). As the measurement data continues to decrease towards the 0.1 range, the algorithm 1260 enters step 1262. At this stage, the mixer actively combines the gains for the 0.1 (A _0.1 ) and 1 (A ₁ ) ranges to smooth the transition to the 0.1 range. Finally, in step 1264, the measurement data is now reliably within the 0.1 range. The mixer provides only the gain (A _0.1 ) associated with the lowest 0.1 measurement data range.

Although FIG. 12B has been described in terms of a decrease in measurement range (increasing gain) from the highest 10 range to the lowest 0.1 measurement data range, it should be understood that FIG. 12B is bidirectional. That is, algorithm 1250 can also proceed if the measurement data is increasing from the range 0.1 to 10 with a corresponding decrease in gain. In that case, algorithm 1260 would follow steps in the reverse order, i.e., 1264-1252.

Figures 13A and 13B show a flow chart illustrating a range change prediction algorithm 1300 that may be performed by the mixer when implementing the algorithms disclosed herein (e.g., 600, 620, 910, 1200, and 1250).

The algorithm 1300 begins by initializing an input signal. In step 1302, the input signal is measured. A first range A is activated for comparison to the input signal in step 1302. The comparison is made in step 1304.

If range A is not desired, algorithm 1300 determines whether the range is too low or too high in step 1306. If the range is too high, the range is decreased in step 1308a. If range A is too low for the measured input signal, the gain for range A is increased in step 1308b. Whether range A is increased or decreased, the next step 1310 waits for any transient effects caused by the gain change to dissipate. Following transient dissipation, algorithm 130 again performs step 1302, measuring the signal and comparing it to the revised gain for range A.

If step 1304 determines that the gain associated with range A is desired for the measured signal, then algorithm 1300 proceeds to step 1312. In step 1312, the algorithm predicts a change from range A to new range B. This "warms up" the electronics associated with the new range B. In step 1314, algorithm 1300 initiates the input based on its assessment of the new range B and the measured input. In step 1316, algorithm 1300 measures the input in both ranges A and B. In step 1318, algorithm 1300 selects the best range of ranges A and B for the measured input to be active (i.e., for use in measuring the input).

The algorithm 1300 then begins the process of determining a switching threshold based on the measured input and the current ranges A and B. In step 1320, the algorithm 1300 determines whether the active range is below the lower range switching threshold. If the active range is below the lower range threshold, the algorithm 1300 performs step 1322 to determine whether the cold or unused range of ranges A and B is within the lower range. If cold range A or B is within the lower range, the algorithm 1300 proceeds to step 1324 to begin blending. If cold range A is not the lower range, the algorithm 1300 sets the cold range to the lower range in step 1326 and then proceeds to step 1324 to begin blending.

If the algorithm 1300 determines in step 1320 that the active range is not below the lower switch threshold, it proceeds to step 1328. In step 1328, the algorithm 1300 determines whether the active range is above the upper switch threshold. If so, the algorithm 1300 proceeds to step 1330 to determine whether the cold or unused range of ranges A and B is the higher range. If cold range A or B is in the higher range, the algorithm 1300 proceeds to step 1324 to start blending. If cold range A is not the higher range, the algorithm 1300 sets the cold range to the higher range in step 1332 and then proceeds to step 1324 to start blending.

If the algorithm 1300 finds that the active range does not fall below the lower switch threshold (step 1320) and also does not exceed the upper switch threshold (step 1328), the algorithm proceeds to step 1334. In step 1334, the algorithm 1300 applies hysteresis to prevent a range change. This is because the measured signal is not within the upper or lower thresholds of the range change. Therefore, any decision to change the range would be based on erroneous noise or glitches in the data. Once hysteresis has been applied, the algorithm proceeds to step 1324 and begins mixing.

In step 1324, the algorithm 1300 begins the steps to start blending. The first step is to make sure the cold range is stable. If the cold range is stable, the system is ready to blend. The algorithm 1300 then proceeds to step 1326 to determine whether to blend the range. If the decision is to blend, the algorithm 1300 blends the range in step 1328 and then provides the blended signal as an output in step 1330. If the decision is not to blend, the algorithm sets the output to the active range in step 1332. If the cold range is not stable, the algorithm 1300 proceeds from step 1324 to step 1332 and sets the output to the active range. After the output is set to the active range in 1332, the signal is then output in step 1330.

Ranging does not have to be performed exclusively by an algorithm. It can also be performed via hardware. Figure 14 shows the parameters input by one such example hardware variant 1400. In 1400, there is a "primary" channel capable of measuring named ranges and a "secondary" channel with less gain. At each ranging update, the percentage of full scale indication on the primary channel can be used to determine the behavior as follows:

Figure 14 illustrates how the range blending algorithm 1400 would behave with respect to different input levels. More specifically, Figure 14 illustrates how the algorithm 1400 would blend (i.e., "mix") different channel gains A and B based on the inputs (i.e., range, range enumeration, input voltage, preamp enable, stage B enable, stage C enable, channel A gain, and channel B gain). The inputs relate to two gain channels A and B and two sample stages B and C. The "range enumeration" is an integer representation of a particular range (i.e., 10V range is "0", 1V range is "1", 100mV range is "2", etc.).

Figure 15 illustrates a measurement signal chain 1500 between an example head unit 1550 and an example measurement pod 1560 that may use variations 400, 500, 700, 800, 1000, 1100 and algorithms 600, 620, 910, 1200, 1250, 1300, and 1400. While Figure 15 illustrates certain aspects of seamless ranging in system 1500, it should be understood that system 1500 may accommodate any variation disclosed herein.

As shown in FIG. 15, the head 1550 includes a measurement channel 1502. In the exemplary case, there are two input measurement channels, one for range A and one for range B, each with its associated ADC. Note that in some variations, each measurement unit 1560 will have an associated configuration 1500 in communication with the head 1550. This means that a variation with three measurement pods 1560 may have six ADCs. It should be understood that any suitable number of measurement channels is possible, depending on the particular measurement and the number of ranges involved, which may be substantially more than two (e.g., three, four, or more). The measurement channel 1502 may be obtained from the measurement pod 1560 via a number of variable amplifiers 1520 and analog filters 1504, as shown in FIG. The gain on the amplifier 1520 may be set as described in the context of gains 1520a-1520c in Figures 10-12. The channel 1502 may be combined 1506 with a range mixed signal 1508 and sent for demodulation 1510 via lock-in. The demodulation may be informed by a reference signal (e.g., Reference (Lock-in) and Reference +90 Degrees (Lock-in) 1512) and may undergo a digital filter 1514 for signal refinement.

The signal can be processed in any number of ways, including DC, AC, or lock-in processing. The range determination can be based on the peak value of the measured sample signal, regardless of other processing being performed for the measurement. This is because the peak value is what would cause amplifier overload. As shown in FIG. 15, the range mixer 1508 may also provide an output regarding range and settings 1516 that is ultimately fed back to the amplifier 1520 and analog filter 1504 to adjust the gain and processing of the measured sample signal for each of ranges A and B. The range mixer 1508 may function as described above in the context of range mixers 410, 510, 710, 1010, and 1110. This process is referred to as continuous measurement ranging and/or range mixing. The purpose is to ensure that there are no glitches or measurement discrepancies that may otherwise occur when the measurement pod 1560 must change its acquisition parameters to adjust for changes in the range of the measured sample signal.

The measurement pod 1560 may further include digital (non-analog) circuitry capable of performing various functions, including analysis, communication of data, command information, power regulation, timing, and communication with external devices. In a variation, the measurement pod 1560 has the ability to deactivate this non-analog circuitry while performing measurements or providing a source signal. Doing so reduces the amount of interference and noise in the signal or measurement. For the same reasons, the digital signals in the measurement pod 1560 may be isolated from the source pod 1560 and head 1550.

Other variations of the system 1500 include any suitable number of heads 1550, source pods, and measurement pods 1560. For example, FIG. 16 shows another exemplary variation 1600 in which the head unit 1550 may have six channels that may support three measurement type pods 1560a and three source type pods 1560b. In this variation, the head 1550 is also shown connected to an optional computer 1602 and three exemplary sampled or devices under test (DUTs) 1570. Again, this configuration is merely exemplary. There is no requirement for an equal number of measurement pods 1950a and source pods 1950b. One source 1950a may, for example, provide excitation signals for all three DUTs 1570.

Although various inventive aspects, concepts, and features of the invention may be described and illustrated herein as embodied in combination in exemplary embodiments, these various aspects, concepts, and features may be used in many alternative embodiments, either individually or in various combinations and subcombinations thereof. Unless expressly excluded herein, all such combinations and subcombinations are intended to be within the scope of the invention. Still further, various alternative embodiments of the various aspects, concepts, and features of the invention may be described herein, such as alternative materials, structures, configurations, methods, circuits, devices and components, software, hardware, control logic, form, compatibility, and function, but such descriptions are not intended to be a complete or exhaustive list of available alternative embodiments, whether currently known or later developed. Those skilled in the art may readily adopt one or more of the inventive aspects, concepts, or features into additional embodiments and uses within the scope of the invention, even if such embodiments are not expressly disclosed herein.

In addition, although some features, concepts, or aspects of the invention may be described herein as being preferred arrangements or methods, such description is not intended to imply that such features are required or necessary unless expressly so described. Still further, exemplary or representative values and ranges may be included to aid in understanding the present disclosure, however, such values and ranges are not intended to be construed in a limiting sense, and are intended to be critical values or ranges only if so expressly described. Still further, exemplary or representative values and ranges may be included to aid in understanding the present disclosure, however, such values and ranges are not intended to be construed in a limiting sense, and are intended to be critical values or ranges only if so expressly described. Parameters identified as "approximately" or "about" a specified value are intended to include both the specified value and values within 10% of the specified value, unless expressly described otherwise. Furthermore, it is to be understood that the drawings accompanying this application may, but need not be, to scale and thus may be understood as teaching various ratios and proportions apparent in the drawings. Also, although various aspects, features, and concepts may be expressly identified herein as being inventive or forming part of the invention, such identification is not intended to be exclusive; rather, there are inventive aspects, concepts, and features that are fully described herein without being expressly identified as such or as part of a specific invention, and the invention may instead be set forth in the appended claims. The description of an exemplary method or process is not limited to the inclusion of every step as required in all cases, nor should the order in which the steps are presented be construed as required or necessary unless expressly so recited.

Claims (21)

1. A measurement system comprising:
a gain chain configured to amplify the analog input signal;
a range selector configured to select a gain between the analog input signal and a plurality of analog-to-digital converter (ADC) outputs from a plurality of ADCs, each ADC output having an output path, the gain of each output path being configured from a plurality of gain stages in the gain chain;
a mixer configured to combine the multiple ADC outputs into a single mixed output;
the plurality of ADCs includes a first ADC and a second ADC;
Combining the plurality of ADC outputs comprises:
Mixed output = αE _first + (1-α)E _second
Implemented in accordance with
In the formula,
E _first is the output of the first ADC;
E _second is the output of the second ADC;
α is a mixing parameter that ranges from 1 to zero in the measurement system. The measurement system of claim 1 comprising three or more ADCs. the gain chain comprises a first portion and a second portion;
2. The measurement system of claim 1, wherein the first portion of the gain chain is connected to the first ADC and the second portion of the gain chain is connected to the second ADC . 4. The measurement system of claim 3 , wherein the range selector selects a gain for the first ADC from the first portion of the gain chain and selects a gain for the second ADC from the second portion of the gain chain. 2. The measurement system of claim 1, wherein each of the multiple gain stages in the gain chain is connected to each of the multiple ADCs via one or more switch banks. the gain stage comprises a first section and a second section;
6. The measurement system of claim 5, wherein the range selector selects the first portion of the multiple gain stages for the first ADC and the second portion of the multiple gain stages for the second ADC by setting switches in the one or more switch banks. the gain stage comprises a first section and a second section;
the range selector comprises first and second multiplexers;
6. The measurement system of claim 5, wherein the first multiplexer selects the first portion and the second multiplexer selects the second portion. 8. The measurement system of claim 7, wherein selecting the first portion includes configuring a gain for the first ADC and selecting the second portion includes configuring a gain for the second ADC . 9. The measurement system of claim 8, wherein configuring a gain for the first ADC and the second ADC comprises configuring the gain according to the analog input signal . the gain of each output path is the same ;
10. The measurement system of claim 1, wherein the mixer averages the output from each output path to reduce noise in the single mixed output. 1. A measurement system comprising:
a gain chain configured to amplify the analog input signal;
a range selector configured to select a gain between the analog input signal and a plurality of analog-to-digital converter (ADC) outputs from a plurality of ADCs, each ADC output having an output path, the gain of each output path being configured from a plurality of gain stages in the gain chain;
A mixer, the mixer comprising:
selecting an output from a first ADC as a first single mixed output when a value of the analog input signal is within a first range;
selecting an output from a second ADC as a second single mixed output when a value of the analog input signal is within a second range;
selecting a combination of the output from the first ADC and the output from the second ADC as a third single mixed output when the value of the analog input signal is between an upper limit of the first range and a lower limit of the second range. The measurement system comprises:
maintaining the second ADC online during a first transition period when the value of the analog input signal is within the first range;
maintaining the first ADC online during a second transition period when the value of the analog input signal is within the second range;
The measurement system of claim 11. During the hysteresis period, the measurement system:
maintaining the first ADC offline;
maintaining said second ADC online;
maintaining the gain of the second ADC constant;
13. The measurement system of claim 12. The measurement system of claim 13 , wherein the hysteresis period is between the first transition period and the second transition period. 1. A measurement system comprising:
a gain chain configured to amplify the analog input signal;
a range selector configured to select a gain between the analog input signal and a plurality of analog-to-digital converter (ADC) outputs from a plurality of ADCs, each ADC output having an output path, the gain of each output path being configured from a plurality of gain stages in the gain chain;
a mixer configured to combine the multiple ADC outputs into a single mixed output;
The plurality of ADC outputs include
Two output paths that can be independently configured into a high range path and a low range path,
the low range path having a first gain for converting the analog input signal;
the high range path has a second gain for converting the analog input signal, the second gain being lower than the first gain;
Two output paths;
a mixing device configured to combine an output of the low range path with an output of the high range path ;
A measurement system comprising: 16. The measurement system of claim 15 , wherein the high range path is connected to a first gain chain and the low range path is connected to a second gain chain. 17. The measurement system of claim 16 , further comprising a selector for selecting a gain stage of the first gain chain for the first gain and for selecting a gain stage of the second gain chain for the second gain. 16. The measurement system of claim 15 , wherein the first and second gains each comprise a gain stage in a gain chain common to the low range path and the high range path. 1. A method comprising:
amplifying an analog input signal using a gain chain;
selecting a gain between the analog input signal and a plurality of analog-to-digital converter (ADC) outputs from a plurality of ADCs, the plurality of ADCs comprising a first ADC and a second ADC, each ADC output having an output path, the gain of each output path being configured from a gain stage in the gain chain;
combining the multiple ADC outputs into a single mixed output according to mixed output=αE _first +(1−α)E _second ;
In the formula,
E _first is the output of the first ADC;
E _second is the output of the second ADC;
α is a mixing parameter that ranges from 1 to zero;
wherein each of the gain stages in the gain chain is connected to each of the plurality of ADCs via one or more switch banks. the gain chain comprises a first portion and a second portion;
20. The method of claim 19 , wherein the first portion is connected to the first ADC and the second portion is connected to the second ADC. configuring two output paths independently into a high range path and a low range path;
applying a first gain from the low range path to convert the analog input signal;
applying a second gain from the high range path to convert the analog input signal, the second gain being lower than the first gain; and
combining an output of said low range path with an output of said high range path ;
20. The method of claim 19 , further comprising:

Images